We hypothesize that amyloid -protein (A) assembly into neurotoxic oligomers and polymers is a seminal neuropathogenetic process in Alzheimer's disease (AD). If so, assembly inhibition or dissociation of existing assemblies could be effective therapeutic approaches. To test our hypothesis, the structural biology of A must be elucidated in detail. What conformers form oligomers? By what mechanism? What are the structures of the oligomers thus formed? What is the relative toxicity of each oligomer species? Many, including ourselves, have striven to correlate structure with measures of biological activity. Recent work has suggested that dimeric or trimeric assemblies are important neurotoxins, but hexameric, nonameric, dodecameric, and larger oligomers also have been shown to be potent neurotoxins. The long-term goal of this proposal is to move past simple quaternary structure determination to elucidation of A monomer secondary and tertiary structure dynamics and the determination of mechanisms of monomer oligomerization. This means eventually understanding the interatomic interactions that control the dynamics, and in doing so, identifying therapeutic targets at atomic resolution. This knowledge-based approach is distinct from, but complementary to, high-throughput screening strategies. Both approaches should be executed to maximize the chances for identifying efficacious, disease-modifying therapeutic agents. We propose here to: (1) elucidate the physical biochemistry of A monomer folding and self-assembly; and (2) establish structure-neurotoxicity relationships of the A assemblies thus formed. To do so, we will chemically synthesize A peptides in which specific amino acids and chemical bonds are altered and then study the conformational dynamics and assembly of these peptides. The positions of these alterations, and the alterations themselves, have been chosen carefully so as to reveal the key structural features of the A molecule that control its assembly into structures that damage or kill neurons. We will identify, isolate, and structurally characterize specific types of assemblies and then determine quantitatively the toxic activity of each assembly by treating primary neurons in culture. The depth of understanding of the structures of the assemblies obtained in the first aim will be unprecedented. Thus the knowledge gained through this structure-activity correlation process is expected to provide the most accurate assessment of which assemblies, and which structures (at atomic resolution) on these assemblies, should be targeted therapeutically. In addition to its contributions to an improved understanding of AD and its treatment, results of the proposed project should have relevance for studies of other neurodegenerative diseases linked to aberrant protein assembly. These include Parkinson's, Huntington's, amyotrophic lateral sclerosis, familial amyloid polyneuropathy, and the prionoses.